Temporal Authority: Why Update Frequency is an AI Trust Signal — Not Just an SEO Signal
AI systems weight content recency as a trust input. A site with high authority that stops updating is not stable — it is decaying. The Temporal Authority Model tracks velocity, drift, and decay across your entire content history.
In traditional SEO, content freshness matters for certain query types — news, current events, time-sensitive searches. In AI retrieval, freshness operates differently. It is not a query modifier but a trust signal. AI systems that have processed your content multiple times across different crawl dates develop a model of your content's stability. Content that changes unpredictably creates semantic drift. Content that has not been updated for months on a topic where information evolves triggers trust decay. SiteNexis models both.
Authority Velocity
Authority velocity is the rate at which your AI-visible authority is growing or declining across consecutive audits. We compute delta on three metrics: Entity Confidence Score, Citation Probability Score, and External Validation signal count. A domain where Entity Confidence increased from 54 to 68 between audits, Citation Probability increased from 61 to 74, and two new sameAs validations resolved has strongly positive velocity. A domain where all three declined has negative velocity — a warning that the content or entity infrastructure is degrading.
●Authority Velocity requires a minimum of two audits to compute. On first audit, SiteNexis returns a baseline record with velocity: null and status: baseline_established. Run a second audit after making improvements to see velocity scores.
Semantic Drift Detection
Semantic drift occurs when a page's content changes meaning over time — shifting from one topic cluster to another without a canonical redirect, or gradually drifting away from its original entity anchor. We detect drift by comparing embedding cosine similarity between page body text snapshots from consecutive audits. A page that was firmly in the "machine trust modeling" topic cluster and has drifted toward "general SEO tips" has a high semantic drift score. Drifted pages confuse the AI's accumulated knowledge model of your domain.
Trust Decay Modeling
Trust decay applies when content ages without update signals. Pages without dateModified schema receive an accelerated decay rate. Pages with schema-confirmed update dates receive a decay slowdown. The decay parameters are configurable — not hardcoded — because decay rates differ by content type. A product page with pricing data decays faster than an evergreen methodology post. A news article becomes stale in days; a technical specification may remain fresh for years. SiteNexis loads decay parameters from a configuration file, not from code.
Update Frequency Classification
We classify every domain into one of four update frequency states: Actively maintained (updated weekly or more), Periodically maintained (updated monthly), Stale (last meaningful update 3+ months ago), or Abandoned (6+ months with no detectable updates). Classification is based on dateModified schema values, HTTP Last-Modified headers, and sitemap change frequency attributes. Stale and Abandoned classifications are negative velocity signals — AI systems that have indexed and cached your content will progressively down-weight it.
Content Freshness Impact Factor
Not all content decays at the same rate. We tag claims containing dates, statistics, version numbers, and pricing as time-sensitive. Time-sensitive claims on stale pages carry a freshness penalty in the Temporal Authority calculation. A page claiming "as of 2023" with no update since then has high freshness penalty. A page claiming "this methodology has three phases" with no time-sensitive data has near-zero freshness penalty. Evergreen structural content is treated differently from data-containing content.